#导入库import pandas as pdimport numpy as npfrom sklearn.preprocessing import Imputer#生成缺失数据df=pd.DataFrame(np.random.randn(6,4),columns=['col1','col2','col3','col4'])df.iloc[1:2,1] = np.nan #增加缺失值df.iloc[4,3] = np.nan #增加缺失值print(df) #打印输出 col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 NaN -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 NaN5 1.002177 0.448844 -0.584634 -1.038151#查看缺失值位置nan_all=df.isnull()print(nan_all) col1 col2 col3 col40 False False False False1 False True False False2 False False False False3 False False False False4 False False False True5 False False False Falsenan_col1=df.isnull().any() #获取含有NA的列print(nan_col1)col1 Falsecol2 Truecol3 Falsecol4 Truedtype: boolnan_col2=df.isnull().all() #获得全部为NA的列print(nan_col2)col1 Falsecol2 Falsecol3 Falsecol4 Falsedtype: bool#丢弃缺失值df2=df.dropna() #直接丢弃含有NA的行纪录print(df2) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896952 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152655 1.002177 0.448844 -0.584634 -1.038151#通过sklearn的数据预处理方法对缺失值进行处理nan_model=Imputer(missing_values='NaN',strategy='mean',axis=0) #建立替换规则：将值为NaN的缺失值以均值做替换nan_result=nan_model.fit_transform(df) #应用模型规则print(nan_result) #打印输出[[-0.97751051 -0.56633185 -0.52993389 1.48969465] [-0.49112788 -0.25284792 -0.81117388 -1.10271738] [ 0.38577678 -0.63882219 0.32595345 -0.24077995] [ 0.93835121 -0.74688892 0.37519957 -0.71526484] [ 1.10341788 0.23895916 -0.45911413 -0.32144373] [ 1.00217657 0.4488442 -0.58463419 -1.03815116]]#使用Pandas做缺失值处理nan_result_pd1 = df.fillna(method='backfill') #用后面的值替换缺失值print(nan_result_pd1) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 -0.638822 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 -1.0381515 1.002177 0.448844 -0.584634 -1.038151nan_result_pd2 = df.fillna(method='bfill',limit=1) #用后面的值替换缺失值，限制每列只能替代一个缺失值print(nan_result_pd2) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 -0.638822 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 -1.0381515 1.002177 0.448844 -0.584634 -1.038151nan_result_df3=df.fillna(method='pad') #用前面的值替换缺失值print(nan_result_df3) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 -0.566332 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 -0.7152655 1.002177 0.448844 -0.584634 -1.038151nan_result_df4=df.fillna(0) #用0替换缺失值print(nan_result_df4) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 0.000000 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 0.0000005 1.002177 0.448844 -0.584634 -1.038151nan_result_df5=df.fillna({'col2':1.1,'col4':1.2}) #用不同值替换不同列的缺失值print(nan_result_df5) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 1.100000 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 1.2000005 1.002177 0.448844 -0.584634 -1.038151nan_result_df6=df.fillna(df.mean()['col2':'col4']) #用各自列的平均数替换缺失值print(nan_result_df6) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 -0.252848 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 -0.3214445 1.002177 0.448844 -0.584634 -1.038151nan_result_df7=df.replace(np.nan,0) #用Pandas的replace替换缺失值print(nan_result_df7) col1 col2 col3 col40 -0.977511 -0.566332 -0.529934 1.4896951 -0.491128 0.000000 -0.811174 -1.1027172 0.385777 -0.638822 0.325953 -0.2407803 0.938351 -0.746889 0.375200 -0.7152654 1.103418 0.238959 -0.459114 0.0000005 1.002177 0.448844 -0.584634 -1.038151